L.B. Foster: The Industrial That Won't Chase Data Centers
Q3 sales rose 21.5%, debt hit a decade low, and a precast bolt-on closed — but mix-driven gross margin compression is the catch.
FSTR · Earnings Call · 2026-11-04
The anti-AI trade
On a call where every industrial CEO is asked about data centers, L.B. Foster's John Kasel went the other way.
We're doing this without a lot of data center work. In fact, less than 2% of our sales is in data centers.
That line matters because the company's own keyword list still flags data centers as a top-ten theme — a reminder that NLP extracts phrases, not intent. The phrase surfaced because an analyst asked about it, and management swatted it away. Meanwhile the global tape shows the AI data center complex down more than 20% over the last 30 days. FSTR isn't riding that wave, and right now that's a feature. The company is a rail and infrastructure maintenance play, not an AI infrastructure proxy.
Organic growth, cash, and the bolt-on they promised
The Q3 numbers were strong. Sales of $168 million rose 21.5%, adjusted EBITDA of $13.1 million rose 15.1%, and operating cash flow of $23.6 million let the company cut debt to $28.7 million — the lowest in more than a decade. Effective net cash That's a genuine balance-sheet transformation, not a footnote. Free cash flow guidance was raised to $20–28 million. The cash came from lower working capital and strength in the Protective Coatings business, which management flagged as a Q4 upside driver. The fresh news is the Dulworth acquisition — an $18 million-revenue precast wastewater bolt-on in Texas that closed October 28. “the Dulworth-based business is expected to generate revenue of approximately $18 million per annum and is expected to provide an EBITDA contribution ... approximately $6.5 million-$7.5 million after the completion of the integration activities” — John Kasel, President and CEO · 2026-11-04 That's a 36–42% EBITDA margin on acquired revenue, which is why the deal matters despite its small size. And it's not a surprise: on the May call, Kasel said “we continue to look at bolt-on type operations and that we will be able to add additional product lines or geographic expansion for us.” — John Kasel, President and Chief Executive Officer · 2026-05-04 Dulworth is exactly that — a bolt-on in the Infrastructure Solutions segment's precast growth platform, expanding geographic reach in the Southeast.
The margin question
The catch is mix. Gross margin fell 200 basis points to 20.5%. “Gross margin declined 200 basis points to 20.5%, primarily reflecting a higher mix of sales from our lower margin Rail Distribution and U.K. businesses.” — Sean Reilly, Chief Financial Officer · 2026-11-04 That's the lower margin Rail Distribution and U.K. businesses diluting the consolidated number. SG&A as a percentage of sales fell 190 basis points to 14.1%, so operating leverage is working — but it's opex leverage, not gross margin leverage. When analyst Julio Romero pressed on the implied Q4 EBITDA margin of 8.5–9.5%, Kasel leaned on exactly that: “a lot of leverage going on right now, Julio. ... It's just the leverage of the overall business. In both segments, we're busy. Our plants are absorbing...” — John Kasel, President and CEO · 2026-11-04 The market will want to see that flow through in Q4, because the last filed quarter showed gross margin at 22.3% and operating margin at 4.4% — decent, but not yet a step-change. Gross margin There's also a backlog nuance. New orders rose 35.7% to $156 million, but that comparison includes the prior-year Summit cancellation. Excluding it, orders were up 16.7%. That Summit order was a prior-year headwind: “So we didn't cancel it. Our customer canceled it.” — John Kasel, President and CEO · 2025-11-03 Backlog of $234.7 million was down 5.1% year-over-year. On the August call, management said “at least 80% we'll be able to execute this year” — John Kasel, President and CEO · 2026-08-10 — and they're delivering on that. But a shrinking backlog means the 2027 setup depends on continued order flow, not just conversion. The cash generation story is real; the revenue visibility story is good but not yet accelerating.
What to watch
Three things. First, Dulworth integration — management assumes no material Q4 impact because construction season is slowing, so the synergy number is a 2027 event. Second, the Q4 margin step-up. If SG&A leverage can't offset the mix drag, the raised sales guide won't translate into the EBITDA range. Third, the data center denial. In a market that has spent two years paying up for AI infrastructure exposure, FSTR is a rare industrial that can say "we're not that" — and its stock, down 7.8% over 90 days but still near its June high, suggests investors are still deciding whether that's a relief or a missed trade.